FSEG LOGO FIRE SAFETY ENGINEERING GROUP The Queen's Anniversary Prize 2002 The British Computer Society IT Awards 2001 The European IST Prize Winner 2003 The Guardian University Awards Winner 2014
The Faculty of Architecture, Computing & Humanities


The subject of this research relates to the generation of data-sets concerning the pre-movement times of evacuees derived from two unannounced fire drills. This research relates to evacuation analysis and the use of evacuation modelling tools such as  buildingEXODUS.

In evacuation analysis one of the most important factors is the pre-movement time of the occupants.  The pre-movement time (sometimes referred to as response time)  reflects the time spent by the occupant before starting to evacuate. It includes the time required to perceive the event and the time required to respond to the event.  In many situations the pre-movement time may be greater than the actual travel time. The pre-movement time is generally dependent on the nature of the occupancy, the state of the occupants, the quality of the management system, the type of alarm system in place, the presence of additional supporting cues (such as the presence of smoke or instructions from a member of staff) and even the time of day can be a factor. Typical pre-movement times can vary from seconds (occupants are awake, trained, familiar with building, alarm systems and procedures) to many minutes (situations where occupants may require assistance such as in hospitals).  To date, this is arguably one of the poorest documented attributes in fire safety engineering.

The subject of this research relates to the generation of data-sets concerning the pre-movement times of evacuees derived from two unannounced fire drills. In both cases  a traditional bell system is used with the support of trained staff.  The structures examined are a listed university building and a private hospital. The data was generated from the performance of a fire drill involving the two structures, with researchers on hand to collect the relevant data.

The first data-set was derived from the evacuation of the Dreadnought building, - located at the University of Greenwich - which took place on Thursday 2nd March 2000. A longitudinal analysis was conducted prior to the evacuation to determine an appropriate and representative time for the evacuation to take place.  The event was an “unannounced fire drill” of which the students and staff were supposed to be unaware except for key members of staff.

The university employed a procedure whereby once the alarm sounded nominated members of staff swept each of the rooms, forcing students to leave their work and belongings and inform them of the route that they should adopt. The 'surprise' status of the evacuation was of vital importance to the usefulness of the data set provided (and of course in testing the performance of this procedure). The results of the evacuation were recorded on video tape for later analysis.

The three storey Dreadnought building is part of the Maritime Greenwich World Heritage Site and since 1999 has been occupied by the University of Greenwich. The University of Greenwich currently uses the Dreadnought building for a variety of purposes including providing library services, housing student computing facilities and a small canteen.  A population of 398 people consisting of both students and staff was recorded evacuating from the structure. The structure included six staircases and eight exits. A team of 15 researchers was positioned around the building to record the data using video and manual observations.  Sixty-two closed circuit cameras were also used to gather data. From the material gathered by the researchers, a systematic procedure of data extraction was defined allowing the data to be extracted in a consistent and coherent manner.

The second evacuation drill involved the Blackheath Hospital on the 31st October 2000. The Blackheath Hospital is 69-bed acute care hospital that serves the private sector (link to  two views of the front of the hospital view1 and view2).  During this time it was not possible to evacuate inpatients (including the bedridden patients), therefore an evacuation of the outpatients and staff was observed. Only three members of the senior management were aware that the evacuation was going to take place and these had no impact on the outcome of the evacuation. The evacuation was therefore unannounced. The hospital employed standard procedures whereby once the alarm sounded members of staff (usually nurses) swept each of the populated areas, informing patients that they had to leave and the route that they should adopt.

Eight researchers were located around the building, in waiting rooms and corridors to collect the pre-movement time of the evacuees. This data was recorded manually or using video equipment. A degree of cross-referencing between the observations of the researchers was possible to confirm the results produced. As in the Dreadnought evacuation, there were two classes of occupant that could easily be distinguished: the staff and the patients.

The two cases have a number of similarities in that to a large degree the students and patients were dependent upon the actions of the members of staff to inform them about the situation. Their reaction to the members of staff and the information provided is reflected in the evacuee’s pre-movement time.

This research involves categorising the pre-movement times observed according to the location and role of the evacuee, enabling comparison within and between the cases.

For more information about evacuation modelling and the EXODUS software visit the EXODUS Web Pages. For a complete listing of EXODUS and evacuation publications visit the FSEG Publications pages

Copyright © 2003- University of Greenwich | Privacy and cookies | Accessibility | Contact the Webmaster